POLS 6482 ADVANCED MULTIVARIATE STATISTICS
First Assignment
Due 10 September 2001


  1. Do assignment #1 from my Probability and Statistics I course at Carnegie-Mellon University:

    Part I: First Homework Assignment

  2. After you have completed the tasks assigned, I want you to create some variables and run a few regressions.

    1. The first thing we are going to do is run a regression using the 1st dimension DW-NOMINATE score as the dependent variable with two independent variables: a dummy or indicator variable for political party (1 = Republican; 0 ¹ Republican); and the 2nd dimension DW-NOMINATE score. To get the dummy variable we have to use the GENR command in EVIEWS. At the prompt, type:

      genr partydum=(party<>200)*0 + (party=200)*1

      This produces a vector "partydum" that is 1 if Republican, 0 if not Republican (remember Bernie Sanders!). To check your work:

      show party partydum

      Which pops up a window showing the two variables. Party is the ICPSR ID code for political party (100 = Democrat; 200 = Republican). Scroll through and check to see if you did everything correctly and then close the window.

      As a further check use the HIST command in EVIEWS. That is:

      HIST Partydum

    2. Now run the regression:

      ls x1 c partydum x2

      Note that ls (or LS) stands for "Least Squares" and c is a reserved word in EVIEWS that is used for the intercept term. Print the Regression Output and hand it in as a Microsoft WORD document (see below for instructions on how to do this).

    3. Now we are going to check what happens when we eliminate the 2nd dimension DW-NOMINATE score and use a dummy variable for South/North (South = 11 states of the Confederacy plus Kentucky and Oklahoma). To do this we again use the GENR command in EVIEWS:

      genr southdum=0
      genr southdum=((state>=40 and state<=49) or state=51 or state=53 or state=54)*1


      The first command initializes our dummy variable to zero and the second command sets it equal to one for the southern states.

      Now, check to be certain that your commands correctly generated the dummy:

      show state southdum

      Now, run the regression:

      ls x1 c partydum southdum

      and print the Regression Output and hand it in (for instructions, see below).

  3. Now we are going to duplicate the above work in Stata.

    1. Our first task is to paste the Excel spreadsheet into Stata. To do this, start Stata. It will look like this:



      Open the data editor by clicking the icon on the toolbar and you will see this:



      Go to Excel, put the spreadsheet on the clipboard, and then paste it into Stata. You should see this:



      Note that the variable names are carried over from Excel.

    2. You must now enter variable definitions into Stata. To do this, double-click on the name bar and a dialog box will come up. For example, for the variable "congress" you will see:



      Leave the "Name:" field the same but type in the appropriate label and click "OK". For example, "Congress Number" would be an appropriate label for the variable congress. Do this for every variable (renaming ser05 and ser07 to something more descriptive is a good idea!). Now exit the editor (note that your variable descriptions now appear in Stata's variables window). Save the worksheet!

    3. In order to preserve our output we need to open a logging file that we can in turn bring up in Epsilon. To do this click the logging file icon on the tool bar and in the dialog box select type "Log(*.log)". Now, enter the command:

      d

      and you will see:



      Bring your log file up in Epsilon and you will see:



      Paste the results of the d command into your homework answer (this must be a Microsoft Word document) as described below.

    4. In Stata enter the command:

      summ

      and paste the results into your homework answer.

    5. The dummy variables partydum and southdum can be created in Stata in the following manner:

      generate partydum=0
      replace partydum=1 if party==200


      generate southdum=0
      replace southdum=1 if state >= 40 & state <= 49
      replace southdum=1 if state==51
      replace southdum=1 if state==53
      replace southdum=1 if state==54


      It is a very good idea to check your work. You can do this by using the browse command in Stata which is the same as the show command in Eviews. In particular:

      browse party partydum

      and

      browse state southdum

      Open the data editor and type in definitions for partydum and southdum. Enter the d

      and summ

      commands and paste the results into your homework answer.

    6. To replicate the regressions we did in Eviews, enter the following commands (note that Stata automatically puts in the intercept term -- C in Eviews):

      regress x1 partydum x2

      and

      regress x1 partydum southdum

      Paste the results into your homework answer.


How To Import EVIEWS Output into Microsoft WORD

  1. In EVIEWS, open the "Print Setup" option under "file".



  2. Now run your analysis -- for example

    LS X1 C PARTYDUM SOUTHDUM



  3. Now click on "Print" and this sends the output to \EVIEWS3\EXAMPLE.TXT.

  4. Go into Epsilon and open EXAMPLE.TXT:



  5. Highlight the table and put it on the clipboard. Inside WORD simply paste the table into your document. It will look bad. Highlight it and then select Courier New Font, Point Size 10, and Paragraph Format Single Space. It should look like this:
    
    
    ============================================================
    Dependent Variable: X1                                                
    Method: Least Squares                                                 
    Date: 09/20/00   Time: 11:21                                          
    Sample: 1 443                                                         
    Included observations: 443                                            
    ============================================================
         Variable      CoefficientStd. Errort-Statistic  Prob.            
    ============================================================
             C          -0.360976   0.010927  -33.03616   0.0000          
         PARTYDUM        0.768851   0.014171   54.25403   0.0000          
         SOUTHDUM        0.089036   0.015317   5.813061   0.0000          
    ============================================================
    R-squared            0.874830    Mean dependent var 0.067472          
    Adjusted R-squared   0.874261    S.D. dependent var 0.417945          
    S.E. of regression   0.148202    Akaike info criter-0.973732          
    Sum squared resid    9.664088    Schwarz criterion -0.946010          
    Log likelihood       218.6816    F-statistic        1537.611          
    Durbin-Watson stat   1.781986    Prob(F-statistic)  0.000000          
    ============================================================
    
  6. DO THIS FOR ALL YOUR OUTPUT FOR THE HOMEWORK ASSIGNMENTS.